1. Field of the Invention
This invention relates generally to the fabrication of giant magnetoresistive (GMR) magnetic field sensors of a “current-perpendicular-to-the-plane” (CPP) configuration and more particularly to such sensors having novel multilayer structures that incorporate magnetic nano-oxide layers.
2. Description of the Related Art
Magnetic read sensors that utilize the giant magnetoresistive (GMR) effect for their operation are generally of the “current-in-the-plane” (CIP) configuration, wherein current is fed into the structure by leads that are laterally disposed to either side of the active sensor region and moves through the structure essentially within the planes of its magnetic layers. Since the operation of GMR sensors depends on resistance variations of the active magnetic layers caused by changes in the relative directions of their magnetic moments, it is important that a substantial portion of the current pass through those layers so that their resistance variations can have a maximal effect. Unfortunately, GMR sensor configurations typically involve layer stacks comprising layers that are electrically conductive but that play no role in providing resistance variations. As a result, portions of the current can be shunted through regions that are ineffective in terms of sensor operation and, thereby, the overall sensitivity of the sensor is adversely affected. The CPP sensor configuration avoids this current shunting problem by disposing its conducting leads vertically above and below the active sensor stack, so that all of the current passes perpendicularly through all of the layers as it goes from the lower to the upper lead. The configuration of the conducting leads is such that the current goes through the leads, front to back, in opposite directions in each lead but perpendicularly to the ABS of the sensor element, then passes vertically through all layers of the sensor and perpendicularly through their interfacial planes. Dykes et al. (U.S. Pat. No. 5,668,688) disclose a CPP sensor that is generally structured in a spin-valve (SV) configuration and achieves an enhanced ΔR/R as compared with a CIP sensor of similar configuration. The CPP sensor of Dykes essentially consists of an SV stack comprising a pinning layer, a ferromagnetic pinned layer, and a ferromagnetic free layer, with the three being sandwiched between conducting leads.
Barr et al. (U.S. Pat. No. 6,198,609) addresses certain current flow problems that persist even in the CPP design. In particular, even though a substantial portion of the current goes through each layer of the sensor element because of the over and under disposition of its leads, a portion of the current can still be shunted along edge paths that define the lateral width of the element. The method taught by Barr et al. forms a CPP sensor element having apertures to guide the current so as to substantially reduce these disadvantageous edge effects.
The magnetic tunnel junction (MTJ) is a device that is usable as a magnetic field sensor or as a memory cell in a magnetic random access memory (MRAM) array. The operating principle of the MTJ is quite similar to that of the CPP sensor formed in a spin valve configuration. In the MTJ device, two ferromagnetic layers are separated by a contiguous insulating tunnel barrier layer. One ferromagnetic layer has its magnetic moment fixed spatially by an antiferromagnetic layer that is interfacially coupled to it. The other ferromagnetic layer, the “free” layer, has its magnetization vector free to move. The relative positions of the two magnetization vectors then controls the amount of tunneling current that can pass through the insulating tunnel barrier layer. In an MRAM array, such an MTJ structure would be “written” by rotating the magnetization direction of its free layer to some given position relative to the magnetically fixed layer. Gallagher et al. (U.S. Pat. No. 5,650,958) provide such a MTJ structure formed with Ni81Fe19 layers as the pinned and free ferromagnetic layers, a Mn50Fe50 layer as the antiferromagnetic layer and Al2O3 layer as the insulating tunnel layer. Dill et al. (U.S. Pat. No. 5,898,548) teach a method of forming a magnetic read head using a similar MTJ element as a read sensor. Nishimura (U.S. Pat. No. 6,111,784) teaches a method of forming an MTJ structure for use as a magnetic thin film memory, wherein the MTJ structure comprises a first magnetic layer, a non-magnetic, partially insulating tunneling layer and a second magnetic layer, the two magnetic layers having different coercivities. Finally, Lubitz, et al. (U.S. Pat. No. 6,171,693) teaches a method of forming a GMR stack having at least two ferromagnetic layers separated from each other by a nonferromagnetic layer, wherein a layer of phase-breaking material such as Ta or a Ta-alloy between the ferromagnetic layer and the nonferromagnetic layer prevents the undesirable growth of large-grained structures in the ferromagnetic layers.
One problem with CPP sensor configurations has already been alluded to above, the undesirable shunting of current along the edges of the active sensor region. Another more general problem of even greater importance is the difficulty of fabricating a CPP sensor element having a resistance within reasonable bounds for practical applications. In this regard, CPP structures formed of metallic multilayers, such as those cited in the patents above, have too low a resistance, whereas MTJ type configurations, having insulating tunneling layers, have too high a resistance. Taking as a figure of merit RA, the product of perpendicular-to-plane sensor resistance, R, and cross-sectional area, A, it is found that metallic multilayers typically have RA between 1 mΩ.μm2 (1 milli-ohm micron squared) and 5 mΩ.μm2, while MTJ type configurations typically have RA=10 Ω.μm2 or more. The RA value of the metallic multilayers can vary to some degree with the materials used for the layers, the layer thicknesses and the number of repeated layers. Nevertheless, for reading high-density magnetic recordings (above 200 Gbit/in2), the thickness of the sensor is limited by the need to resolve magnetic flux transitions, so it is not possible to increase RA meaningfully by increasing thickness. For an area, A, within usable value of about 0.01 μm2, the CPP resistance is about 0.1 Ω, which is too low for practical purposes. MTJ's have also been considered as possible sensor structures, since large MR amplitudes of up to 40% at room temperature have been reported. In these junctions, as in the magnetic multilayers, the perpendicular resistance, R, varies inversely with the area of the junction, A. Evaluation of the signal-to-noise ratio in MTJ read heads has shown that such heads can compete with CIP sensor heads only if the AR product can be reduced to below 5 Ω.μm2. Such low resistance is difficult to attain in MTJ structures. Since the resistance of such junctions varies exponentially with junction thickness, an Alumina tunnel layer (such as that in Gallagher et al., above) would have to have a thickness of less than 5 angstroms to achieve the requisite RA value. Such a thin layer would introduce the problems of pinholes or general reliability over typical usage periods.
Therefore a need arises for a structure having a value of RA that is intermediate between that of metallic multilayered CPP configurations and MTJ type configurations.